The page tree is not always fully loaded when exported to PDF

The page tree (jsTree) is using a Web Worker to parse the JSON input and create the tree nodes. This is running in a background thread, asynchronously from the main execution thread of the web page. There's no easy or generic way that we could intercept these Web Workers and detect when their work is done (they send messages but these messages are custom, specific to each Web Worker). So the only option for now is to modify the page tree to manually delay the page ready until the tree is fully loaded.